Transcript Sub-seasonal to seasonal Prediction Project An element of the WWRP Background  Several operational centres are now producing sub-seasonal forecasts. There is a.

Sub-seasonal to seasonal Prediction Project
An element of the WWRP
Background
 Several operational centres are now producing sub-seasonal forecasts.
There is a need to fill the gap between medium-range and seasonal
forecasting and link the activities of WCRP and WWRP.
 The WMO Commission of Atmospheric Sciences (CAS) requested at its 15th
session (Nov. 2009) that WCRP, WWRP and THORPEX set up an appropriate
collaborative structure for sub-seasonal prediction.
 A WCRP/WWRP/THORPEX workshop was held at Exeter (1-3 December
2010).
www.wcrp-climate.org/documents/CAPABILITIES-IN-SUB-SEASONAL-TOSEASONAL PREDICTION-FINAL.pdf
Planning Group

The creation of this group follows a main recommendation
from the WWRP/THORPEX/WCRP workshop at the UK Met
Office (1-3 December 2010).
 The planning group was established in 2011
Sponsors: WCRP-WWRP-THORPEX
 Kick-off meeting: 2-3 December 2011
An Implementation plan has been written
Main Goals
The first task of the group was to prepare an implementation plan giving high
priority to:
– The establishment of collaboration and co-ordination between operational
centres undertaking sub-seasonal prediction to ensure when possible
consistency between operational approaches to enable the production of
data bases of operational sub-seasonal predictions to support the
application of standard verification procedures and a wide-ranging program
of research.
– Facilitating the wide-spread research use of the data collected for the CHFP
(and its associate projects), TIGGE and YOTC for research.
– Sponsorship of a few international research activities
– The establishment of a series of regular workshops on sub-seasonal
prediction
Subseasonal to Seasonal Prediction Planning group
Use of sub-seasonal forecasts in applications
Growing, and urgent, requirement for the employment of sub-seasonal
predictions for a wide range of societal and economic applications which
include:
• Warnings of the likelihood of severe high impact weather (droughts,
flooding, wind storms etc.) to help protect life and property
•Humanitarian Planning and Response to disasters
•Agriculture particularly in developing countries — e.g. wheat and rice
production
•Disease planning/control — e.g. malaria, dengue and meningitis
•River-flow — for flood prediction, hydroelectric power generation and
reservoir management for example
Use of sub-seasonal forecasts in applications
• Weather and climate span a continuum of time scales, and forecast
information with different lead times are relevant to different sorts of
decisions and early-warning
• In agriculture, for example, a seasonal forecast might inform a cropplanting choice, while sub-monthly forecasts could help irrigation
scheduling, pesticide/fertilizer application: both can make a cropping
calendar dynamic.
• In situations where seasonal forecasts are already in use, sub-seasonal
ones could be used as updates, such as for end-of-season crop yields.
• Sub-seasonal forecasts may play an especially important role where
initial conditions and intraseasonal oscillation is strong, while seasonal
predictability is weak, such as the Indian summer monsoon.
Opportunity to use information on
multiple time scales
Red Cross - IRI example
Monsoon Onset and Rice Planting in Java, Indonesia
Calendars: Rice-planting area in Indramayu, Java
Source: Boer et al. (2004)
Planting Area (ha)
Rainfall (mm)
rice
Cropping Pattern
rice
Fallow
Start of planting
changes from time to
time, in planting
season 97/98, start of
planting delayed 1
month due to delay
onset of rainfall,
increasing drought
risk for the second
crop, except in LaNina years
Application aspects of SSI prediction: User-relevant needs
• Availability of long hindcast histories are needed to
develop and test regression-based “MOS” and
tailoring models, and for skill estimation which is
critical to applications.
• Daily data is needed, especially for a few key variables
including precipitation and near-surface temperature
and windspeed.
• Issues of open data access to enable uptake
Bridging the gap between Climate prediction and NWP
• A particularly difficult time range: Is it an atmospheric initial
condition problem as medium-range forecasting or is it a
boundary condition problem as seasonal forecasting?
•
Some sources of predictability in the sub-seasonal time scale:
– The Madden Julian Oscillation
– Sea surface temperature/Sea ice
– Snow cover
– Soil moisture
– Stratospheric Initial conditions
Impact of soil moisture
Koster et al, GRL 2010
12
Stratospheric influence on the troposphere?
Baldwin and Dunkerton, 2001
1. Impact of the MJO on Extratropics
F i gu r e 12: V er t i cal l y av er aged an om al ou s h eat i n g r at e f or ( a) E x p 1; an d ( b ) E x p 2. T h e
cont ou r i nt er val i s 0.5 C d ay - 1 . T h e zer o cont ou r i s n ot p l ot t ed , an d cont ou r s w i t h n egat i v e
val u es ar e d ash ed .
eat i n g r at e f or ( a) E x p 1; an d ( b ) E x p 2. T h e
nt ou r i s n ot p l ot t ed , an d cont ou r s w i t h n egat i v e
Figure 13: 500 hPa geopot ent ial height response averaged bet ween day 6 and 10 (left ) and
bet ween day 11 and 15 (right ) for Exp1 (t op) and Exp2 (bot t om). T he cont our int erval is
Lin et al, MWR 2010
See also
Simmons et al JAS 1983
Ting and Sardeshmukh JAS 1993
Figure 13: 500 hPa geopot ent ial height response averaged bet ween day 6 and 10 (left ) and
15 m. Cont ours wit h negat ive values are dashed.
bet ween day 11 and 15 (right ) for Exp1 (t op) and Exp2 (bot t om). T he cont our int erval is
15 m. Cont ours wit h negat ive values are dashed.
43
43
Impact of the MJO on weather regimes
Cassou (2008)
Scientific issues
– Identify sources of predictability at the sub-seasonal time-range
– Prediction of the MJO and its impacts in numerical models
– Teleconnections - forecasts of opportunity
– Monsoon prediction
– Rainfall predictability and extreme events
– Polar prediction and sea-ice
– Stratospheric processes
Modelling issues
– Role of resolution
– Role of Ocean-atmosphere coupling
– Systematic errors
– Initialisation strategies for sub-seasonal prediction
– Ensemble generation
– Spread/skill relationship
– Design of forecast systems
– Verification
Recommendations for coordinated research activities
–
–
–
–
–
–
–
–
–
–
–
–
–
–
Define a common set of common methodologies and metrics to validate models
Identify potential sources of predictability and their representation in models
Identify forecast “windows of opportunity”
Investigate prediction of the onset and cessation of rainy season
Investigate the modulation of extreme events by the MISO
Investigate the prediction of sea-ice and its impact
Set up demonstration projects
Investigate how best to initialise models, to diagnose error growth
Assess the extent to which coupled data assimilation can improve forecast skill
Identify the impact of increased resolution on sub-seasonal forecasts
Evaluate the impact of ocean-atmosphere coupling
Quantify the advantages and disadvantages of burst vs lagged ensemble generation
Develop and compare methodologies for re-calibration of sub-seasonal forecasts.
Quantify the skill of a multi-model ensemble compared to that from a single model.
Some research topics can be coordinated with the appropriate working groups (e.g.
Polar prediction, project, MJO Task Force, GASS...)
Sub-seasonal forecast database
A main recommendations from the Exeter meeting in 2010 was the establishment of a data
bases of operational sub-seasonal predictions. Over the past years, a few multi-model
databases have been set up:






TIGGE (WWRP/THORPEX) : Medium-range forecasts (day 0-14) from 10 operational centres are collected 2 days
behind real-time. Servers: NCAR/ECMWF/CMA.
CHFP (WCRP) data archive: Operational and non-operational seasonal forecast hindcasts from 18 centres.
ET-ELRF (CBS): Real-time seasonal forecasts archived at KMA/NCEP. Available only to WMO members and only the
multi-model combination is available. Limited number of fields available.
Field experiments: real time data for a short period of time
Multi-model operational seasonal forecasts: EUROSIP, APEC Climate Center (Busan)
Hindcast experiments: ENSEMBLES, ISO experiment (APCC) ….
Most of these databases were not designed to investigate the sub-seasonal predictability and do not
include the current operational sub-seasonal forecasting systems. There is therefore a need to
establish a new database which would include operational sub-seasonal forecasts. This database
would be very useful to address most of the recommendations of coordinated research. For some
recommendations, specific coordinated experiments may be needed.
Sub-seasonal forecast database
• Numerical models have shown significant
improvements in sub-seasonal prediction over the
past years (e.g. MJO).
• 10 years ago, only a couple of operational centres
were producing sub-seasonal forecasts. Over the
past years, a few GPCs have set sub-seasonal
forecasting systems.
Examples of improvements in MJO prediction
MJO Bivariate Correlation
Day the MJO bivariate 0.5
Correlation reaches 0.6
32
30
28
26
Forecast Day
24
22
20
18
16
14
12
10
2002
2003
2004
2005
2006
2007
YEAR
2008
2009
2010
2011
Examples of improvements in NAO prediction
MJO Bivariate Correlation
NAO Index Anomaly
Correlation for day 19-25
0.5
0.6
Forecast Day
0.5
0.4
2002
2003
2004
2005
2006
2007
YEAR
2008
2009
2010
2011
Simulation of the impact of the MJO on the NAO
Lin et al, 2010
Sub-seasonal real-time Operational Forecasts
Timerange
Resol.
Ens.
Size
Freq.
Hcsts
Hcst
length
Hcst Freq Hcst Size
ECMWF
D 0-32
T639/319L62
51
2/week
On the fly
Past 18y
weekly
5
UKMO
D 0-60
N96L85
4
daily
On the fly
1989-2003
4/month
3
NCEP
D 0-60
N126L64
16
daily
Fix
1999-2010
daily
4
EC
D 0-35
0.6x0.6L40
21
weekly
On the fly
Past 15y
weekly
4
CAWCR
D 0-120
T47L17
33
weekly
Fix
1989-2010
3/month
33
JMA
D 0-34
T159L60
50
weekly
Fix
1979-2009
3/month
5
KMA
D 0-30
T106L21
20
3/month
Fix
1979-2010
3/month
10
CMA
D 0-45
T63L16
40
6/month
Fix
1982-now
monthly
48
CPTEC
D 0-30
T126L28
1
daily
No
-
-
-
Met.Fr
D 0-60
T63L91
41
monthly
Fix
1981-2005
monthly
11
SAWS
D 0-60
T42L19
6
monthly
Fix
1981-2001
monthly
6
HMCR
D 0-60
1.1x1.4 L28
10
monthly
Fix
1979-2003
monthly
10
Proposal for a sub-seasonal database
• At least 6 centres produce sub-seasonal forecasts every Thursday:
ECMWF, JMA, NCEP, UKMO, CAWCR , EC
• Archive daily means of real-time forecasts + hindcasts.
• Proposed list of archived variables is based on TIGGE + ocean variables and
stratospheric levels (total of about 73 fields)
• Real-time forecasts 3 weeks behind real-time
• Archive the variables in a 1.5x1.5 degree grid or lower once a week.
• Use TIGGE protocol (GRIB2) for archiving the data. The data could also be
archived in NETCDF for WCRP community.
• Make the database also available in netcdf to attract the WCRP community
• Use of the first 2 months of the CHFP seasonal and climate forecasting
systems to compare with the archive (above). Need for daily or
weekly/pentads archive.
Data volume
Hypothesis:
- 1.5x1.5 degree or less
- 73 variables
- All centres are archiving all the fields.
Total cost (real-time + hindcasts from the 12 GPCs) is estimated to:
- 15TB for the first year
- About 7 TB per year in the following years.
This would represent less than 10% of the TIGGE archiving cost (about 180
TB/year at ECMWF). The fixed imposed resolution should keep the cost about
constant from year to year.
The choice of TIGGE protocols and the limited data volume should make it
easier centres like ECMWF to accept to host this dataset.
Demonstration projects
A few case studies to demonstrate that using sub-seasonal predictions
could be of benefit to society.
Cases studies could include:
•
•
•
Pakistan floods (2010) concurrent with the Russian heat wave
Australian floods (2011)
European Cold spell (2011)
At least one of the demonstration projects should be in real-time, which is
often the best way to foster collaborations between the research and
application communities.
The models could be archived near real-time during a limited period of
time with additional fields being archived. The period chosen could
coincide with test bed studies from other projects (e.g. polar project).
Example : Pakistan Floods (2010)
28
30°N
30°N
Verification period: 26-07Verification period: 26-07-2
30°N
30°N
Sub-seasonal Prediction of Pakistan Floods (2010)
20°N
ensemble size = 51 ,climate size = 90
20°N
20°N
20°N
ensem ble size = 51
ensem ble size = 51 ,
Shaded areas signif
Shaded areas signific
Contours at
Contours at 1
20°N
Shaded areas significant at 10%
level
Contours at 1% level
10°N
10°N
10°N
10°N
10°N
Precip anomalies : 26 July– 01 August 2010
40°E
60°E
40°E
40°E
80°E
FORECAST 22-07-2010:
DAY 5-11
ANALYSIS
40°E
40°E 60°E
60°E
80°E
80°E
60°E
60°E
80°E
80°E
FORECAST
22-07-2010:
DAY
5-11
FORECAST
15-07-2010:
DAY
40°E
60°E
80°E
FORECAST
22-07-2010:
DAY
5-1112-18
40°E
60°E
80°E
40°E
60°E
FORECAST 15-07
FORECAST60°E
15-07
40°E
40°E
80°E
40°N
40°N
40°N
40°N and ECMWF VarEPS-Monthly Forecasting
40°N
Analysis
System
40°N
40°N
40°N
<-90mm
Precipitation anomaly
30°N
Verification period: 26-07-2010/TO/01-08-2010 30°N30°N
30°N
40°N
40°N
40°N
Analysis and ECMWF VarEPS-Monthly40°N
Forecasting
System
Precipitation anomaly
30°N
30°N
60°E
30°N
30°N
Verification period: 26-07-2010/TO/01-08-2010
30°N30°N
30°N
30°N
-90..-60
20°N
ensemble size = 51 ,climate size = 90
20°N
Shaded areas significant at 10% level
20°N
20°N
20°N
20°N
20°N
ensemble size = 51 ,climate20°N
size = 90
20°N
Shaded areas
Contours at 1% level
10°N
10°N
60°E
60°E
40°E
60°E
80°E
40°N40°E
40°N
60°E
40°E
60°E
80°E
80°E
80°E
40°N
30°N
20°N
20°N
30°N
20°N
30°N
20°N 20°N
20°N
20°N
10°N
10°N
10°N
20°N
20°N
10°N
10°N
10°N
10°N
40°E
40°E
60°E
60°E
80°E
10°N
80°E
10°N
FORECAST 08-07-2010: DAY 19-25
FORECAST40°E
01-07-2010: DAY
26-32 80°E
60°E
40°E
60°E
80°E
40°E
40°E
60°E
80°E
FORECAST 08-07-2010:
DAY
19-25
FORECAST
15-07-2010: DAY 12-18
40°E
60°E 40°E
80°E
60°E
80°E
40°E
60°E
FORECAST
08-07-2010:
DAY80°E
19-25
80°E
40°N40°E
40°N
60°E
80°E
<-90mm
40°N
40°N
60°E
40°E
60°E
FORECAST
01-07
-10..
0
40°E
60°E
FORECAST
01-07
40°E
40°N
40°N
30°N 30°N
30°N
30°N
60°E
-30..-10
60°E
80°E
FORECAST
01-07-2010:
DAY 26-32
40°N
30°N
10°N
40°E
40°N 40°N
40°N
30°N
-60..-30
10°N
10°N
80°E
FORECASTFORECAST
15-07-2010:22-07-2010:
DAY 12-18 DAY 5-11
40°E
10°N
10°N
FORECAST 08-07-2010: DAY 19-25
40°N
10°N
10°N
10°N
40°E
20°N
20°N
Contours at 1% level
10°N
10°N
40°E
20°N
significant at 10%
level
40°N
60°E
0.. 10
<-90mm
10.. 30
-90..-60
40°N
40°N
30°N
30°N
30°N
-90..-60
20°N
30°N
30°N
20°N
20°N
-60..-30
10°N
20°N
10°N
20°N
10°N
-30..-10
-10..
0
40°E
40°E
10°N
60°E
60°E
40°E
20°N
20°N
10°N
10°N
10°N
80°E
80°E 60°E
FORECAST
01-07-2010: DAY 26-32
80°E
40°E
0.. 10
30°N
30°N
60°E
29
80°E
30°N
20°N
30.. 60
10°N
-60..-30
-30..-10
60.. 90
40°E
> 90mm
40°E
-10.. 0
60°E
60°E
0.. 10
Linkages
• Global Framework for Climate Services
• CLIVAR and GEWEX including regional panels and
WGNE
• Year of Tropical Convection
• CBS
• Verification working groups (SVS-LRF and JWGFVR)
• World Bank
Main recommendations
• The
establishment of a project Steering group
• The establishment of a project office
•The establishment of a multi-model data base consisting of ensembles of subseasonal (up to 60
days) forecasts and re-forecasts
• A major research activity on evaluating the potential predictability of subseasonal events,
including identifying windows of opportunity for increased forecast skill.
•A series of science workshops on subseasonal to seasonal prediction.
•Appropriate demonstration projects based on some recent extreme events and their impacts
This project will require 5 years, after which the opportunity for a 5 year extension will
be considered.